基于小波互能量交叉的复杂背景中红外小目标检测方法

Wavelet Analysis Based and Mutual Energy Combination Method Detection Algorithm for Infrared Small Target in Complex Background

研究复杂背景下红外图像小目标检测问题,提出一种基于小波分析互能量交叉处理的目标检测方法。通过定义水平、垂直方向上的小波互能量交叉函数,在两上方向上有效地增强目标、抑制背景干扰。采用一种简单有效的局部直方图统计方法进行目标粗分割,综合两个方向目标粗检测结果通过能量最大判决进行目标定位,并采用双窗不相似度量检测目标尺寸。实验结果表明,该方法对地面背景、海面背景的红外弱小目标检测具有一定的普遍适用性,具有单帧检测率高、实现速度快的特点。

The small target detection method of infrared imagery in complex background of sea and sky in studied,and an algorithm based on wavelet analysis and mutual energy combination is presented.Wavelet Mutual Energy Combination (WMEC) functions are defined in horizontal and vertical directions. Targets are enhanced and disturbances of background are weakened efficiently by WMEC processing at the two directions. Targets are coarsely segmented with thresholds which are set automatically by a method of part gray histogram analysis. Results of coarse segmentation in the two directions are integrated, and targets are located by means of maximum energy judgment. The size of targets is decided by comparing the dissimilarity between a couple of windows. Experiment results indicate that the method can be appropriate in both kinds background of land and sea. It is with high correct detection ratio from a single image, and high speed to realize.